Adaptive Model Predictive Control of a Quadrotor

摘要

This work explores the application of Adaptive Model Predictive Control (AMPC) to quadrotor altitude control. Model Predictive Control (MPC) is a very powerful method of Advanced Control, utilizing an implicit model to make output predictions which are in turn used in computing control action. Dynamic Matrix Control (DMC) was the first MPC technique to be implemented, and is still very common in process control especially for Chemical plants. Due to changes in operating points, the implicit model in MPC typically becomes insufficient, costing the quality of the controller. This work proposes Adaptation (based on the Recursive Least Squares (RLS) algorithm), for online system identification to take changes in operating points into account when computing the DMC control action. Overall, Adaptive DMC is investigated as to whether Adaptive DMC is capable of improved quadrotor control or tracking.